Skip to content

Benchmark

LiveBench tablereformat

Scores are percentages.

Models measured
51
Scores recorded
51
Sources
1
IQ weight
0.89

Score distribution

98%
Mean
98%
Median
2.3%
Std dev
96.1%
10th percentile
100%
90th percentile
86.3%
Min
86.3%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.89 for chat models, measured over 51 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1DeepSeek V4 Flash 0731· LiveBench data100%
2DeepSeek V4 Pro· LiveBench data100%
3GPT 5.2high· LiveBench data100%
4GPT 5.4xhigh· LiveBench data100%
5GPT 5.5xhigh· LiveBench data100%
6GPT-5.2 Codex· LiveBench data100%
7GPT-5.4 minixhigh· LiveBench data100%
8GPT-6.1 Solxhigh· LiveBench data100%
9Grok 4.5· LiveBench data100%
10Grok 4.6· LiveBench data100%
11Grok 4.7xhigh· LiveBench data100%
12Grok Build 0.1· LiveBench data100%
13Inkling FP4xhigh· LiveBench data100%
14Mistral Large 4high· LiveBench data100%
15Muse Spark 1.1xhigh· LiveBench data100%
16Muse Spark 1.2xhigh· LiveBench data100%
17Muse Spark 1.3xhigh· LiveBench data100%
18Claude Haiku 5.5xhigh· LiveBench data98%
19Claude Opus 4.564k, high· LiveBench data98%
20Claude Opus 4.6auto, high· LiveBench data98%
21Claude Opus 4.7xhigh· LiveBench data98%
22Claude Opus 5.5xhigh· LiveBench data98%
23Claude Sonnet 5xhigh· LiveBench data98%
24Claude Sonnet 5.5xhigh· LiveBench data98%
25DeepSeek V4 Flash· LiveBench data98%
26DeepSeek V4 Flash Vision Exp· LiveBench data98%
27DeepSeek V4 Pro 0813· LiveBench data98%
28Gemini 3.1 Pro Preview· LiveBench data98%
29Gemini 3.5 Flash Litehigh· LiveBench data98%
30Gemini 3.6 Flashhigh· LiveBench data98%
31Gemini 3.7 Flashhigh· LiveBench data98%
32GPT 5.4 Nanoxhigh· LiveBench data98%
33Grok 4.3· LiveBench data98%
34Kimi K2.6thinking· LiveBench data98%
35Kimi K3· LiveBench data98%
36MiniMax M3· LiveBench data98%
37Nemotron 3 Ultra· LiveBench data98%
38Qwen3.6 27B· LiveBench data98%
39Qwen3.6 Plus· LiveBench data98%
40Qwen3.7 Max· LiveBench data98%
41Qwen3.8 Max· LiveBench data98%
42Claude Sonnet 4.6auto, medium· LiveBench data96.1%
43Gemini 3.5 Flashhigh· LiveBench data96.1%
44Gemini 3.8 Flashhigh· LiveBench data96.1%
45GLM 5.2· LiveBench data96.1%
46GLM 5.3· LiveBench data96.1%
47GLM 5.3 Flash· LiveBench data96.1%
48Qwen 3.8 27B· LiveBench data96.1%
49Qwen 3.8 Flash Next· LiveBench data96.1%
50Kimi K2.7 Code· LiveBench data92.2%

Showing 1–50 of 51 results.